A comparative analysis of machine learning techniques and key insights for cardiovascular disease prediction
Hossain, Jobayel (2024)
2024
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2024060521137
https://urn.fi/URN:NBN:fi:amk-2024060521137
Tiivistelmä
Heart disease, or cardiovascular disease (CVD) is a prevalent and deadly condition globally. The purpose of this thesis was to develop and evaluate a robust system applying classification-based machine learning methods to predict heart disease using a comprehensive dataset of patient health information, including age, sex, chest pain type, blood pressure, cholesterol levels, and maximum heart rate. The goal was to improve the accuracy and reliability of heart disease predictions, thereby assisting healthcare providers.
The study was developed by significant patterns and correlations involving data preprocessing, exploratory data analysis, and visualization techniques. Classification algorithms such as Logistic Regression, Decision Tree, Support Vector Machine, and Gaussian Naive Bayes were implemented and achieved cross-validation accuracies of 86.62%, 80.51%, 86.77%, 86.48%, and 85.89%, respectively. A combined method using a Voting Classifier method that achieved a prediction accuracy of 87.83%. Feature importance analysis provided insights into key predictors of heart disease, aiding medical professionals in the decision-making process.
The findings suggest that standard models provide a robust framework for early detection and effective management of heart disease. Healthcare institutions can start using these models right away to improve diagnostic processes. Further research is required to integrate real-time data and expand the model to incorporate additional health metrics. Overall, this research underscores the transformative potential of advanced diagnostic models in healthcare, offering extensive opportunities for both medical professionals and patients in the management and treatment of heart disease.
The study was developed by significant patterns and correlations involving data preprocessing, exploratory data analysis, and visualization techniques. Classification algorithms such as Logistic Regression, Decision Tree, Support Vector Machine, and Gaussian Naive Bayes were implemented and achieved cross-validation accuracies of 86.62%, 80.51%, 86.77%, 86.48%, and 85.89%, respectively. A combined method using a Voting Classifier method that achieved a prediction accuracy of 87.83%. Feature importance analysis provided insights into key predictors of heart disease, aiding medical professionals in the decision-making process.
The findings suggest that standard models provide a robust framework for early detection and effective management of heart disease. Healthcare institutions can start using these models right away to improve diagnostic processes. Further research is required to integrate real-time data and expand the model to incorporate additional health metrics. Overall, this research underscores the transformative potential of advanced diagnostic models in healthcare, offering extensive opportunities for both medical professionals and patients in the management and treatment of heart disease.